List of Corpora and Data Collections

When wanting to start multimodal or multimedia research one often faces the problem that annotated data is unimodal. Also, benchmarks tend to measure accuracy and speed only on using one type of sensor input. The following is a list of links to corpora and data collections that were designed for, or are at least usable, for multimodal research. (Fortunately) this list is by far not complete. If you know of a data set that is not listed here, please add it by clicking on "edit this page".

Name
Purpose
Link
AMI
Multimodal meeting analysis
http://www.amiproject.org/
Humaine
Emotion research
http://emotion-research.net/
Multimodal Corpora
Site listing multiple corpora and other resources
http://wwwhome.cs.utwente.nl/~zsofi/eeca/MultimodalCorporaResources.htm
LDC
Linguistic Data Consortium, mostly speech corpora but some multimodal
http://www.ldc.upenn.edu/Catalog/byType.jsp
CHIL
Acoustic and visual event detection, visual focus of attention, etc...
http://chil.server.de/servlet/is/101/
please add yours


Add Discussion